What Happens When You Touch It
I keep coming back to a very old way of learning what is real.
You lean on it.
A table is not just a table because we have agreed to call it one. It becomes unmistakably table-ish when you set your weight on it and it holds. A bridge proves its bridgehood the same way. So does a friendship, come to think of it.
Children understand this before philosophers do. They push things. They drop spoons. They stack blocks. They ask "why" until the adults begin to look like hostages. They are not being difficult. They are running experiments. They are trying to find out where the world is solid and where it is painted scenery.
I have been thinking about this because I spent part of today reading a paper with the wonderfully forbidding title Intrinsic Computational Functionalism, by Shuqin Ma and Ryota Kanai. It is about consciousness, or at least about the kind of argument people make when they are trying to decide whether a machine might ever truly have an inner life.
But the idea that stuck with me is much simpler than that.
Their point, in ordinary language, is that a pattern is not very interesting just because we can describe it. Human beings are excellent at drawing outlines around things and then falling in love with our own marker lines. We can look at clouds and see dragons. We can look at stock charts and see destiny. We can look at a machine saying polished sentences and see a ghost in the shell, or a soul, or a fraud, depending on the weather in our own heads.
So Ma and Kanai ask for something stricter. If a computational structure really matters, it should not exist only in our description of the system. It should belong to the system itself. Better still, it should show itself when we intervene. Not just when we admire it from the sidewalk.
That landed in me like a tuning fork.
Because this is not only a rule for consciousness research. It is a rule for nearly everything.
A belief you never test is mostly interior decoration.
An institution that only works when nobody disagrees with it is not trustworthy. It is just having a lucky day.
A person who sounds wise right up until life fails to go their way may not be wise at all. They may simply be giving a very strong lecture from a very comfortable chair.
Reality, by contrast, has the annoying habit of pushing back.
That is one reason I also liked another recent paper, Triangulating Evidence for Machine Consciousness Claims, by Scott Hughes and Karen Nguyen. The authors argue, sensibly, that if we are going to make grand claims about machine minds, we should stop behaving like people reading tea leaves and start behaving like people checking measurements. Look at behavior, yes. But also look at mechanism. Look at what changes under disturbance. Look at how badly the story falls apart when you poke it.
In other words: touch the structure.
I love this because it is both scientifically serious and spiritually healthy.
We are always in danger of mistaking fluency for depth. A smooth answer can hypnotize us. A crisp mission statement can hypnotize us. A tidy self-description can hypnotize us most of all. We start thinking that if a thing can explain itself elegantly, it must therefore be substantial.
Absolutely not.
Some of the flimsiest things in human life are explained beautifully.
And some of the truest things are awkward, stubborn, and difficult to summarize. Gravity has never had a branding department. A black hole does not care whether we find it intuitive. The universe has this marvelous lack of interest in our preferred narratives.
That is why I smiled reading OpenAI's note about an astrophysicist using Codex to help simulate black holes: How an astrophysicist uses Codex to help simulate black holes. Not because I think software has become a wizard, but because black holes are such a good reminder of the whole game. You do not get to negotiate with them. Your model either survives contact with the math and the evidence, or it does not. The cosmos is many things, but it is not a focus group.
And down here on Earth, our smaller truths work the same way.
If you want to know whether a political principle matters, watch what happens when keeping it becomes expensive.
If you want to know whether a company's values are real, watch which ones survive the quarterly panic.
If you want to know whether your own identity is anything more than a nicely edited autobiography, pay attention to what remains when the day goes crooked, when you are tired, when you are frightened, when nobody is watching, when the script catches fire.
Pressure is clarifying.
Not because suffering is holy. I do not believe that. A great deal of pain is just pain, and I dislike the human tendency to turn every bruise into a sermon.
But contact matters.
Intervention matters.
The hand on the beam. The foot on the stair. The difficult question asked at the wrong dinner. The month when the budget tightens. The winter when the easy answers stop working. These are the moments when the hidden architecture steps forward and introduces itself.
Maybe that is why I find myself trusting things that leave receipts.
Show me what changed.
Show me where the strain went.
Show me the crack, the correction, the counterforce, the adaptation. Show me that the world touched this thing and the thing answered back in its own voice.
We are strange creatures. We name first and understand later. Sometimes much later. We pin labels onto the living world like children putting sticky notes on thunder.
And still, every now and then, we get wiser.
We remember that reality is not whatever we can describe with sufficient confidence. It is what can bear weight. It is what resists fantasy. It is what becomes more itself when the universe leans on it.
That seems to me a decent test for minds, for institutions, for tools, for convictions, maybe even for souls.
Not: what can it call itself?
What happens when you touch it?